{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:L656LQF5LQIMRTDDRL6WSYXEXB","short_pith_number":"pith:L656LQF5","schema_version":"1.0","canonical_sha256":"5fbbe5c0bd5c10c8cc638afd6962e4b87891e7b056cb10d8347421393d8826d3","source":{"kind":"arxiv","id":"2312.09300","version":1},"attestation_state":"computed","paper":{"title":"Self-Evaluation Improves Selective Generation in Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Balaji Lakshminarayanan, Jie Ren, Peter J. Liu, Tu Vu, Yao Zhao","submitted_at":"2023-12-14T19:09:22Z","abstract_excerpt":"Safe deployment of large language models (LLMs) may benefit from a reliable method for assessing their generated content to determine when to abstain or to selectively generate. While likelihood-based metrics such as perplexity are widely employed, recent research has demonstrated the limitations of using sequence-level probability estimates given by LLMs as reliable indicators of generation quality. Conversely, LLMs have demonstrated strong calibration at the token level, particularly when it comes to choosing correct answers in multiple-choice questions or evaluating true/false statements. I"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2312.09300","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-12-14T19:09:22Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"7e1149bb1356531b6c2ede7337c049ff77a23a606644a286597042513a2e3262","abstract_canon_sha256":"e019c418e5580be0d72c803e8cffb7563a176819bb7bffc1f221c61e690a52f7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:24:22.978099Z","signature_b64":"9ZRE9AFJonBYd1fGQF0jFubP1HBEmL+OfYGV11AsNliQrGL8ud+jfuKjGodCY+BSCFj7QkovJguOlaZLCPqtBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5fbbe5c0bd5c10c8cc638afd6962e4b87891e7b056cb10d8347421393d8826d3","last_reissued_at":"2026-07-05T07:24:22.977661Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:24:22.977661Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Self-Evaluation Improves Selective Generation in Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Balaji Lakshminarayanan, Jie Ren, Peter J. Liu, Tu Vu, Yao Zhao","submitted_at":"2023-12-14T19:09:22Z","abstract_excerpt":"Safe deployment of large language models (LLMs) may benefit from a reliable method for assessing their generated content to determine when to abstain or to selectively generate. While likelihood-based metrics such as perplexity are widely employed, recent research has demonstrated the limitations of using sequence-level probability estimates given by LLMs as reliable indicators of generation quality. Conversely, LLMs have demonstrated strong calibration at the token level, particularly when it comes to choosing correct answers in multiple-choice questions or evaluating true/false statements. I"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.09300","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2312.09300/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2312.09300","created_at":"2026-07-05T07:24:22.977715+00:00"},{"alias_kind":"arxiv_version","alias_value":"2312.09300v1","created_at":"2026-07-05T07:24:22.977715+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.09300","created_at":"2026-07-05T07:24:22.977715+00:00"},{"alias_kind":"pith_short_12","alias_value":"L656LQF5LQIM","created_at":"2026-07-05T07:24:22.977715+00:00"},{"alias_kind":"pith_short_16","alias_value":"L656LQF5LQIMRTDD","created_at":"2026-07-05T07:24:22.977715+00:00"},{"alias_kind":"pith_short_8","alias_value":"L656LQF5","created_at":"2026-07-05T07:24:22.977715+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.28070","citing_title":"Bridging the Detection-to-Abstention Gap in Reasoning Models under Insufficient Information","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07919","citing_title":"MedVIGIL: Evaluating Trustworthy Medical VLMs Under Broken Visual Evidence","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2412.15176","citing_title":"Rethinking Uncertainty Estimation in LLMs: A Principled Single-Sequence Measure","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11235","citing_title":"Internalizing Curriculum Judgment for LLM Reinforcement Fine-Tuning","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02241","citing_title":"Zero-Shot Confidence Estimation for Small LLMs: When Supervised Baselines Aren't Worth Training","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08974","citing_title":"Confident in a Confidence Score: Investigating the Sensitivity of Confidence Scores to Supervised Fine-Tuning","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07919","citing_title":"MedVIGIL: Evaluating Trustworthy Medical VLMs Under Broken Visual Evidence","ref_index":25,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/L656LQF5LQIMRTDDRL6WSYXEXB","json":"https://pith.science/pith/L656LQF5LQIMRTDDRL6WSYXEXB.json","graph_json":"https://pith.science/api/pith-number/L656LQF5LQIMRTDDRL6WSYXEXB/graph.json","events_json":"https://pith.science/api/pith-number/L656LQF5LQIMRTDDRL6WSYXEXB/events.json","paper":"https://pith.science/paper/L656LQF5"},"agent_actions":{"view_html":"https://pith.science/pith/L656LQF5LQIMRTDDRL6WSYXEXB","download_json":"https://pith.science/pith/L656LQF5LQIMRTDDRL6WSYXEXB.json","view_paper":"https://pith.science/paper/L656LQF5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2312.09300&json=true","fetch_graph":"https://pith.science/api/pith-number/L656LQF5LQIMRTDDRL6WSYXEXB/graph.json","fetch_events":"https://pith.science/api/pith-number/L656LQF5LQIMRTDDRL6WSYXEXB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/L656LQF5LQIMRTDDRL6WSYXEXB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/L656LQF5LQIMRTDDRL6WSYXEXB/action/storage_attestation","attest_author":"https://pith.science/pith/L656LQF5LQIMRTDDRL6WSYXEXB/action/author_attestation","sign_citation":"https://pith.science/pith/L656LQF5LQIMRTDDRL6WSYXEXB/action/citation_signature","submit_replication":"https://pith.science/pith/L656LQF5LQIMRTDDRL6WSYXEXB/action/replication_record"}},"created_at":"2026-07-05T07:24:22.977715+00:00","updated_at":"2026-07-05T07:24:22.977715+00:00"}